Session: DFMLC-08-01/DAC-20-01: Modeling and Optimization for Sustainable Design and Manufacturing

Paper Number: 68167

Start Time: August 19, 03:20 PM

68167 - Selection Method of Molding Condition for Self-Adhesive Products Using Only Bamboo Fibers Extracted With a Machining Center Based on Bayesian Optimization 

The negative impacts of bamboo, which grows widely throughout Japan, on the ecosystem and soil are currently a serious issue. Studies to find effective uses of excessive bamboo are ongoing. In this study, a method for manufacturing a self-adhesive compact material by pressurizing and heating 100% bamboo fiber after extraction using a machining center, without losing the composition of bamboo, is proposed. This molded body is made of a 100% natural material and has a low environmental impact even when landfilled, contributing to the growth of bamboo for use as a raw material.

In this study, the formation conditions for maximizing the tensile properties when forming platelets were investigated. First, an analysis of the variance was conducted using the orthogonal array test design to identify the factors affecting the objective variable and determine the contribution of each factor. In addition, estimates of the objective variable for the molding conditions were obtained from the Gaussian process regression model in a Bayesian optimization. Furthermore, the maximum values of the characteristic values of the molding conditions were compared between the analysis of variance and the Gaussian process regression model. The results indicate that the Gaussian process regression model is useful tensile property evaluations.

Presenting Author: Daigo Tauchi Doshisha University

Authors:

Daigo Tauchi Doshisha University
Toshiki Hirogaki Doshisha University
Eiichi Aoyama Doshisha University
Keiji Ogawa Ryukoku University
Hiromichi Nobe MIFUJI-KIKAI, INC.